In vitro testing strategy for assessing the skin sensitizing potential of "difficult to test" cosmetic ingredients.

Before placing a new cosmetic ingredient on the market, manufacturers must establish its safety profile, in particular assessing the skin sensitization potential, which is a mandatory requirement for topical applications. Since the ban on animal testing in Europe, and its extension to many parts of the world, a battery of in vitro tests covering the key steps of the Adverse Outcome Pathway (AOP) for skin sensitization is recommended. To date, three in vitro methods are validated in the OECD guidelines (442C, 442D, 442E), and many others are under validation by OECD (2019) and ECVAM. However, there is still no official strategy. Some industrial manufacturers have proposed in vitro strategies with good predictivity, but their studies were mainly based on the testing of simple and "easy to test" substances. This work therefore focused on "difficult to test" ingredients with particular physicochemical properties (i.e. poorly water-soluble components) or with particular intrinsic properties placing them outside the applicability domains of most in vitro models (irritants or cytotoxic like surfactants, complex substances). Furthermore a particular focus was made on weak to moderate sensitizers. The objective was to develop a robust, quick and straightforward testing strategy enabling the evaluation of the skin sensitization potential of "difficult to test" ingredients. In this context, four in vitro test models were used: three validated methods and the Sens-Is® assay, currently in the work plan of the OECD, chosen for its ability to overcome solubility issues and to discriminate irritants from sensitizers. 25 ingredients with particular physicochemical properties were evaluated, chosen among positive or negative sensitizers according to in vivo data (M&K and/or LLNA). Such ingredients, including cleansers, solubilizers, emulsifiers, emollients, active ingredients, preservatives, and antioxidants are indeed essential constituents of cosmetic and dermopharmaceutical formulations. The results analysis on each in vitro test demonstrated that the DPRA model was the less predictive on the chosen ingredients, resulting especially in many false negative responses compared to animal studies, or being unsuited to the mode of action of the selected ingredients. On the contrary, the Sens-Is® assay revealed a real capability to discriminate sensitizers from non-sensitizers. The two other models, KeratinoSensTM and h-CLAT, showed a lower ability to classify the materials correctly than in previously published studies, linked to the particular physicochemical and intrinsic properties of the chosen ingredients and the applicability domains of these in vitro tests. The KeratinoSensTM model tended to overestimate the sensitization potential of the tested ingredients, whereas the h-CLAT model tended to underestimate the sensitizers. Based on these results a new sequential testing strategy was set up combining 1 to 3 models to cover the main key events of the skin sensitization AOP. Sens-Is® model, assessing the first two AOP Key Events with consideration of the ingredient dermal penetration, is chosen as a starting point. The approach is completed, depending on the first response, by the h-CLAT model, assessing Key Event 3, and then potentially KeratinoSensTM assessing Key Event 2, but with a more direct application mode. This new testing strategy increases the accuracy to 88% on the selected ingredients and minimizes the risk of a false negative conclusion, which is crucial from the perspective of the ingredients' users and cosmetic consumers.

Proposal of GHS sub-categorization criteria for LLNA: BrdU-ELISA (OECD TG442B).

The Globally Harmonized System of Classification and Labelling of Chemicals (GHS) is a hazard classification and communication system for providing information on the safe handling of chemicals worldwide. While the GHS provides sub-categorization criteria for sensitizers when using the guinea pig maximization test/Buehler test (OECD TG406) and the standard radioisotopic LLNA (OECD TG429), the sub-categorization criteria for LLNA: BrdU-ELISA (OECD TG442B) are not currently provided. In this study, we re-analyzed the existing data of 32 sensitizers classified in the 1A or 1B categories of the GHS, and attempted to determine optimal criteria for GHS sub-categorization using LLNA: BrdU-ELISA. Consequently, the optimal criterion for the GHS sub-categorization was determined to be 6% when using EC1.6, showing the correct outcomes (%) for GHS 1A and GHS 1B category chemicals were 92.3 and 84.2 for all 32 chemicals, respectively. When excluding 2-mercaptobenzothiazole which may cause strain specific low response in this assay system, the correct outcomes (%) for GHS 1A chemicals was 100. Further work would be necessary, but the GHS sub-categorization criteria proposed in this study might be promising when using LLNA: BrdU-ELISA to provide information on the skin sensitization potency category of chemicals.

Prediction of the skin sensitising potential and potency of compounds via mechanism-based binary and ternary classification models.

Skin sensitisation, one of the most frequent forms of human immune toxicity, is authenticated to be a significant endpoint in the field of drug discovery and cosmetics. Due to the drawbacks of traditional animal testing methods, in silico methods have advanced to study skin sensitisation. In this study, mechanism-based binary and ternary classification models were constructed with a comprehensive data set. 1007 compounds were collected to develop five series of local and global models based on mechanisms. In each series, compounds were classified into five groups according to EC3 values, and applied as training sets, test sets and external validation sets. For each of the five series, 81 binary classification models and 81 ternary classification models were acquired via 9 molecular fingerprints and 9 machine learning methods using a novel KNIME workflow. Meanwhile, the applicability domains for the best 10 models were figured out to certify the rationality of prediction effect. In addition, 8 toxic substructures probably causing skin sensitisation were identified to speculate whether a compound is a skin sensitiser. The mechanism-based prediction models and the toxic substructures can be applied to predict the skin sensitising potential and potency of compounds.

A defined approach for predicting skin sensitisation hazard and potency based on the guided integration of in silico, in chemico and in vitro data using exclusion criteria.

A decision tree-based defined approach (DA) has been designed using exclusion criteria based on applicability domain knowledge of in chemico/in vitro information sources covering key events 1-3 in the skin sensitisation adverse outcome pathway and an in silico tool predicting the adverse outcome (Derek Nexus). The hypothesis is that using exclusion criteria to de-prioritise less applicable assays and/or in silico outcomes produces a rational, transparent, and reliable DA for the prediction of skin sensitisation potential. Five exclusion criteria have been established: Derek Nexus reasoning level, Derek Nexus negative prediction, metabolism, lipophilicity, and lysine-reactivity. These are used to prioritise the most suitable information sources for a given chemical and results from which are used in a '2 out of 3' approach to provide a prediction of hazard. A potency category (and corresponding GHS classification) is then assigned using a k-Nearest Neighbours model containing human and LLNA data. The DA correctly identified the hazard (sensitiser/non-sensitiser) for 85% and 86% of a dataset with reference LLNA and human data. The correct potency category was identified for 59% and 68% of chemicals, and the GHS classification accurately predicted for 73% and 76% with reference LLNA and human data, respectively.

Evaluation of radioisotopic and non-radioisotopic versions of local lymph node assays for subcategorization of skin sensitizers compliant to UN GHS rev 4.

Recently UN GHS has introduced the sub-categorization of skin sensitizers for which ECt (concentration estimated to induce stimulation index above threshold) of the murine local lymph node assay (LLNA) is used as criteria. Non-radioisotopic variants of LLNA, LLNA: DA, LLNA: BrdU-ELISA, LNCC and LLNA: BrdU-FCM were developed yet their utilities for potency sub-categorization are not established. Here we assessed the agreement of LLNA variants with LLNA or human data in potency sub-categorization for 22 reference substances of OECD TG429. Concordance of sub-categorization with LLNA was highest for LLNA: BrdU-FCM(91%, κ = 0.833, weighted kappa) followed by LLNA: BrdU-ELISA (82%, κ = 0.744) and LLNA: DA (73%, κ = 0.656) whereas LNCC only showed a modest association (64%, κ = 0.441). With human data, LLNA agreed best (77%) followed by LLNA: DA and LLNA: BrdU-FCM(73%), LLNA: BrdU-ELISA (68%) and LNCC(55%). Bland-Altman plot revealed that ECt's of LLNA variants largely agreed with LLNA where most values fell within 95% limit of agreement. Correlation between ECt's of LLNA and LLNA variants were high except for LNCC(pair-wise with LLNA, LLNA: DA, r = 0.848, LLNA: BrdU-ELISA, r = 0.744, LLNA: BrdU-FCM, r=0.786, and LNCC, r = 0.561 by Pearson). Collectively, these results demonstrated that LLNA variants exhibit performance comparable to LLNA in the potency sub-categorization although additional substances shall be analyzed in the future.

Assessment of skin sensitization under REACH: A case report on vehicle choice in the LLNA and its crucial role preventing false positive results.

In the EU, chemicals with a production or import volume in quantities of one metric ton per year or more have to be tested for skin sensitizing properties under the REACH regulation. The murine Local Lymph Node Assay (LLNA) and its modifications are widely used to fulfil the data requirement, as it is currently considered the first-choice method for in vivo testing to cover this endpoint. This manuscript describes a case study highlighting the importance of understanding the chemistry of the test material during testing for 'skin sensitization' of MCDA (mixture of 2,4- and 2,6-diamino-methylcyclohexane) with particular focus on the vehicle used. While the BrdU-ELISA modification of the LLNA using acetone/olive oil (AOO) as vehicle revealed expectable positive results. However, the concentration control analysis unexpectedly revealed an instability of MCDA in the vehicle AOO. Further studies on the reactivity showed MCDA to rapidly react with AOO under formation of various imine structures, which might have caused the positive LLNA result. The repetition of the LLNA using propylene glycol (PG) as vehicle did not confirm the positive results of the LLNA using AOO. Finally, a classification of MCDA as skin sensitizer according to the Globally Harmonized System (GHS) was not justified.

Consensus of classification trees for skin sensitisation hazard prediction.

Since March 2013, it is no longer possible to market in the European Union (EU) cosmetics containing new ingredients tested on animals. Although several in silico alternatives are available and achievements have been made in the development and regulatory adoption of skin sensitisation non-animal tests, there is not yet a generally accepted approach for skin sensitisation assessment that would fully substitute the need for animal testing. The aim of this work was to build a defined approach (i.e. a predictive model based on readouts from various information sources that uses a fixed procedure for generating a prediction) for skin sensitisation hazard prediction (sensitiser/non-sensitiser) using Local Lymph Node Assay (LLNA) results as reference classifications. To derive the model, we built a dataset with high quality data from in chemico (DPRA) and in vitro (KeratinoSens™ and h-CLAT) methods, and it was complemented with predictions from several software packages. The modelling exercise showed that skin sensitisation hazard was better predicted by classification trees based on in silico predictions. The defined approach consists of a consensus of two classification trees that are based on descriptors that account for protein reactivity and structural features. The model showed an accuracy of 0.93, sensitivity of 0.98, and specificity of 0.85 for 269 chemicals. In addition, the defined approach provides a measure of confidence associated to the prediction.

A data-based exploration of the adverse outcome pathway for skin sensitization points to the necessary requirements for its prediction with alternative methods.

This paper presents new data-based analyses on the ability of alternative methods to predict the skin sensitization potential of chemicals. It appears that skin sensitization, as shown in humans and rodents, can be predicted with good accuracy both with in vitro assays and QSAR approaches. The accuracy is about the same: 85-90%. Given that every biological measure has inherent uncertainty, this performance is quite remarkable. Overall, there is a good correlation between human data and experimental in vivo systems, except for sensitizers of intermediate potency. This uncertainty/variability is probably the reason why alternative methods are quite efficient in predicting both strong and non-sensitizers, but not the intermediate potency sensitizers. A detailed analysis of the predictivity of the individual approaches shows that the biological in vitro assays have limited added value in respect to the in chemico/QSAR ones, and suggests that the primary interaction with proteins is the rate-limiting step of the entire process. This confirms evidence from other fields (e.g., carcinogenicity, QSAR) indicating that successful predictive models are based on the parameterization of a few mechanistic features/events, whereas the consideration of all events supposedly involved in a toxicity pathway contributes to increase the uncertainty of the predictions.

Analysis of the Local Lymph Node Assay (LLNA) variability for assessing the prediction of skin sensitisation potential and potency of chemicals with non-animal approaches.

The knowledge of the biological mechanisms leading to the induction of skin sensitisation has favoured in recent years the development of alternative non-animal methods. During the formal validation process, results from the Local Lymph Node Assay (LLNA) are generally used as reference data to assess the predictive capacity of the non-animal tests. This study reports an analysis of the variability of the LLNA for a set of chemicals for which multiple studies are available and considers three hazard classification schemes: POS/NEG, GHS/CLP and ECETOC. As the type of vehicle used in a LLNA study is known to influence to some extent the results, two analyses were performed: considering the solvent used to test the chemicals and without considering the solvent. The results show that the number of discordant classifications increases when a chemical is tested in more than one solvent. Moreover, it can be concluded that study results leading to classification in the strongest classes (1A and EXT) seem to be more reliable than those in the weakest classes. This study highlights the importance of considering the variability of the reference data when evaluating non-animal tests.

Evaluation of toxic endpoints for a set of cosmetic ingredients with CAESAR models.

The randomly selected set of 558 chemicals from Cosmetic inventory was studied with internet accessible program package CAESAR. Four toxic endpoints were considered: mutagenicity, carcinogenicity, developmental toxicity and skin sensitization. The CAESAR program provides beside the predictions comprehensive information on applicability domain and the similarity between the considered compound and the compounds from model's training set. This information was used to implement for clustering and classification of chemicals. As the technique the Self Organizing Maps was applied. This technique also enables us to define to each cluster the cluster indicator, i.e., the characteristic compound, which is considered as a representative for a cluster.

Method for detecting the reactivity of chemicals towards peptides as an alternative test method for assessing skin sensitization potential.

Cosmetics are normally composed of various ingredients. Some cosmetic ingredients can act as chemical haptens reacting toward proteins or peptides of human skin and they can provoke an immunologic reaction, called as skin sensitization. This haptenation process is very important step of inducing skin sensitization and evaluating the sensitizing potentials of cosmetic ingredients is very important for consumer safety. Therefore, animal alternative methods focusing on monitoring haptenation potential are undergoing vigorous research. To examine the further usefulness of spectrophotometric methods to monitor reactivity of chemicals toward peptides for cosmetic ingredients. Forty chemicals (25 sensitizers and 15 non-sensitizers) were reacted with 2 synthetic peptides, e.g., the cysteine peptides (Ac-RFAACAA-COOH) with free thiol group and the lysine peptides (Ac-RFAAKAA-COOH) with free amine group. Unreacted peptides can be detected after incubating with 5,5'-dithiobis-2-nitrobenzoic acid or fluorescamine™ as detection reagents for free thiol and amine group, respectively. Chemicals were categorized as sensitizers when they induced more than 10% depletion of cysteine peptides or more than 30% depletion of lysine peptides. The sensitivity, specificity, and accuracy were 80.0%, 86.7% and 82.5%, respectively. These results demonstrate that spectrophotometric methods can be an easy, fast, and high-throughput screening tools predicting the skin sensitization potential of chemical including cosmetic ingredient.

Use of the human monocytic leukemia THP-1 cell line and co-incubation with microsomes to identify and differentiate hapten and prohapten sensitizers.

Consumer and medical products can contain leachable chemical allergens which can cause skin sensitization. Recent efforts have been directed at the development of non-animal based tests such as in vitro cell activation assays for the identification of skin sensitizers. Prohapten identification by in vitro assays is still problematic due to the lack of prohapten bioactivation. The present study evaluated the effect of hapten and prohapten exposure on cell surface markers expression (CD86, CD54 and CD40) in the human monocytic leukemia, THP-1, cell line. Upregulation of activation and costimulatory markers are key events in the allergic sensitization process and have been reported to serve as indicators of skin sensitization. Cells were exposed to the prohaptens benzo(a)pyrene (BaP), 7,12-dimethylbenz(a)anthracene (DMBA), carvone oxime (COx), cinnamic alcohol (CA) and isoeugenol (IEG) at concentrations ranging from 1 to 10 µM for 24 and 48 h. The direct-binding haptens dinitrochlorobenzene (DNCB), benzoquinone (BQ), hydroxylethyl acrylate (HEA) and benzylbromide (BB) were used as positive controls. Cells were also exposed to the irritants sodium dodecyl sulfate (SDS) and sulfanilamide (SFA). Bioactivation of prohaptens was achieved by adding aroclor-induced rat liver microsomes (S9) to the cell cultures. Consistent upregulation of surface expressions of CD86, CD54 (ICAM-1) and CD40 was observed in THP-1 cells treated with direct-acting haptens (±S9) or prohapten (+S9). Upregulation of these markers was not observed after exposure to skin irritants or prohaptens in the absence of exogenously added S9. In conclusion, modification of in vitro cell culture assays to include co-incubation with microsomes enhances identification of prohaptens and allows them to be clearly distinguished from direct-binding haptens.

Prediction of skin sensitization potential using D-optimal design and GA-kNN classification methods.

Modelling of skin sensitization data of 255 diverse compounds and 450 calculated descriptors was performed to develop global predictive classification models that are applicable to whole chemical space. With this aim, we employed two automated procedures, (a) D-optimal design to select optimal members of the training and test sets and (b) k-Nearest Neighbour classification (kNN) method along with Genetic Algorithms (GA-kNN Classification) to select significant and independent descriptors in order to build the models. This methodology helped us to derive multiple models, M1-M5, that are stable and robust. The best among them, model M1 (CCR(train) = 84.3%, CCR(test) = 87.2% and CCR(ext) = 80.4%), is based on six neighbours and nine descriptors and further suggests that: (a) it is stable and robust and performs better than the reported models in literature, and (b) the combination of D-optimal design and GA-kNN classification approach is a very promising approach. Consensus prediction based on the models M1-M5 improved the CCR of training, test and external validation datasets by 3.8%, 4.45% and 3.85%, respectively, over M1. From the analysis of the physical meaning of the selected descriptors, it is inferred that the skin sensitization potential of small organic compounds can be accurately predicted using calculated descriptors that code for the following fundamental properties: (i) lipophilicity, (ii) atomic polarizability, (iii) shape, (iii) electrostatic interactions, and (iv) chemical reactivity.

SuperHapten: a comprehensive database for small immunogenic compounds.

The immune system protects organisms from foreign proteins, peptide epitopes and a multitude of chemical compounds. Among these, haptens are small molecules, eliciting an immune response when conjugated with carrier molecules. Known haptens are xenobiotics or natural compounds, which can induce a number of autoimmune diseases like contact dermatitis or asthma. Furthermore, haptens are utilized in the development of biosensors, immunomodulators and new vaccines. Although hapten-induced allergies account for 6-10% of all adverse drug effects, the understanding of the correlation between structural and haptenic properties is rather fragmentary. We have developed a manually curated hapten database, SuperHapten, integrating information from literature and web resources. The current version of the database compiles 2D/3D structures, physicochemical properties and references for about 7500 haptens and 25,000 synonyms. The commercial availability is documented for about 6300 haptens and 450 related antibodies, enabling experimental approaches on cross-reactivity. The haptens are classified regarding their origin: pesticides, herbicides, insecticides, drugs, natural compounds, etc. Queries allow identification of haptens and associated antibodies according to functional class, carrier protein, chemical scaffold, composition or structural similarity. SuperHapten is available online at http://bioinformatics.charite.de/superhapten.

Haptens, prohaptens and prehaptens, or electrophiles and proelectrophiles.

It is argued that the term 'hapten', and derived terms such as 'pro-hapten' and 'pre-hapten' are ambiguous and unnecessary. It is proposed that their use be abandoned. Instead, when considering the chemical basis of skin sensitization, it is preferable to classify compounds according to the chemical reaction mechanisms by which they can modify proteins.

HaptenDB: a comprehensive database of haptens, carrier proteins and anti-hapten antibodies.

UNLABELLED: The key requirement for successful immunochemical assay is the availability of antibodies with high specificity and desired affinity. Small molecules, when used as haptens, are not immunogenic. However, on conjugating with carrier molecule they elicit antibody response. The production of anti-hapten antibodies of desired specificity largely depends on the hapten design (preserving greatly the chemical structure and spatial conformation of target compound), selection of the appropriate carrier protein and the conjugation method. This manuscript describes a curated database HaptenDB, where information is collected from published literature and web resources. The current version of the database has 2021 entries for 1087 haptens and 25 carrier proteins, where each entry provides comprehensive details about (1) nature of the hapten, (2) 2D and 3D structures of haptens, (3) carrier proteins, (4) coupling method, (5) method of anti-hapten antibody production, (6) assay method (used for characterization) and (7) specificities of antibodies. The current version of HaptenDB covers a wide array of haptens including pesticides, herbicides, insecticides, drugs, vitamins, steroids, hormones, toxins, dyes, explosives, etc. It provides internal and external links to various databases/resources to obtain further information about the nature of haptens, carriers and respective antibodies. For structure similarity comparison of haptens, the database also integrates tools like JME Editor and JMOL for sketching, displaying and manipulating hapten 2D/3D structures online. So the database would be of great help in identifying functional group(s) in smaller molecules using antibodies as well as for the development of immunodiagnostics/therapeutics by providing data and procedures available so far for the generation of specific or cross-reactive antibodies. AVAILABILITY: HaptenDB is available on http://www.imtech.res.in/raghava/haptendb/ and http://bioinformatics.uams.edu/raghava/haptendb/ (Mirror site).

Phthalic anhydride allergy: development and characterization of optimized hapten-carrier conjugates for improved diagnosis.

BACKGROUND: At present the diagnosis of IgE-mediated hypersensitivity to phthalic anhydride (PA) is based on conjugates that are not characterized or standardized. The aim of this study was to develop optimized and molecularly characterized PA conjugates that can be used to improve the diagnosis of PA-allergy. METHODS: The PA conjugates were synthesized and the number of haptens bound on a carrier protein was estimated by matrix-assisted laser desorption/ionization time of light (MALDI-TOF) mass spectrometry. The ability of conjugates to bind IgE and IgG antibodies was measured by enzyme-linked immunosorbent assay (ELISA). Reactivity of the conjugates in vivo was evaluated by skin prick testing. RESULTS: The most active IgE-binding conjugates had a PA : HSA molar ratio of 80 : 1. In the optimal conjugates the average numbers of PA haptens per carrier molecule of human serum albumin (HSA) were 14-16. In ELISA, all 13 patients and none of the 20 controls had IgE antibodies to optimized PA conjugate. The sensitivity and specificity of the ELISA was comparable to commercial CAP RAST. PA conjugates elicited positive test results in skin prick testing showing that conjugates are immunologically active also in vivo. CONCLUSIONS: These results indicate that optimized and molecularly characterized PA-HSA conjugates can be used both in vitro and in vivo assays to improve the diagnosis of PA allergy.

A novel lectin in rabbit serum binds H type 1, H type 2 and N-acetyl lactosamine structures.

A novel lectin (RSL) which recognizes blood group H type 1 and type 2 (Fuc alpha 1-->2Gal beta 1-->3/4GlcNAc beta-R), and N-acetyllactosamine (Gal beta 1-->4GlcNAc beta-R) was purified from rabbit serum using affinity chromatography on Synsorb H type 2 beads, gel filtration and preparative polyacrylamide gel electrophoresis. The lectin agglutinated human O type red cells, and the hemagglutination reaction was inhibited by H type 1 and type 2 haptens, N-acetyllactosamine and human salivas from secretor individuals. The molecular weight of the lectin was estimated to be approximate 650,000 and 65,000 on Sephacryl S-400 gel filtration and SDS-polyacrylamide gel electrophoresis, respectively.